Previous studies had suggested that group cohesion is related both to efficient group process and positive treatment outcome. This study was designed to explore whether manipulation of interactions among incarcerated addicts would help overcome racial and countercultural resistances and increase cohesion. The study was conducted in a correctional institution DAP (drug abuse program) unit, a specialized live-in program for offenders with drug histories, in which half the offenders were housed in 10-or 11-man cubicles segregated along therapy group lines. Hypotheses were that the cubicle group would be more cohesive and also have greater interracial tolerance. Two questionnaires were given to 79 DAP residents to test therapy group cohesion and interracial tolerance. Thirty-nine of those tested lived in the cubicle section, and 26 of the 39 were black. Results indicated that, as predicted, inmates exposed to the cubicle system displayed significantly more cohesion than did those not exposed to this system, suggesting that combining programmatic and architectural variables can benefit the group treatment process. Results concerning the cubical system's impact on racial bias were less promising, although whites exposed to cubicle living displayed more acceptance of blacks than the control group, suggesting that the cubicle system can increase racial acceptance. Findings also showed that inmates not exposed to cubicle living displayed a significant favoring of therapy group members, implying factors aside from cubicle living may foster therapy group cohesion. The study indicates that intertwining of programmatic and architectural factors might be useful in the treatment process, although further research is necessary. A total of 19 references are listed.